GitHub Actions down again today
Source: Hacker News
Incident Summary
On May 12 2026, between 13:41 and 17:43 UTC, several GitHub services experienced processing delays. For the Code Scanning service, 53 % of check runs took over 15 minutes to complete. Notifications averaged a 22‑minute delivery time, and Slack integration webhooks averaged 20 minutes. The delays were caused by replication lag from an internal database migration, which reduced worker capacity for the high rate of job enqueues.
The impact was mitigated by scaling processing workers, and all services returned to normal after the mitigation. GitHub is creating dedicated worker pools for high‑usage shared queues to help prevent similar incidents in the future.
Timeline of Updates
- May 12 13:41 UTC – Incident began.
- May 12 15:13 UTC – Investigating reports of degraded performance for CodeQL.
- May 12 15:42 UTC – CodeQL actions experiencing delays; actions may remain pending or fail due to timeout.
- May 12 15:44 UTC – Continuing investigation of CodeQL actions workflows; also seeing delays for notifications, webhooks, and Slack integration.
- May 12 16:18 UTC – Webhooks experiencing degraded performance; investigation ongoing.
- May 12 16:28 UTC – Most delays linked to a queuing service; scaling out workers; early signs of recovery for some services.
- May 12 16:29 UTC – Webhooks operating normally.
- May 12 16:59 UTC – Webhooks fully recovered; work continues on other services.
- May 12 17:43 UTC – CodeQL fully recovered; all services have fully recovered.
Resolved – All services returned to normal processing times after scaling the workers.
We continue to monitor the system and will provide updates as needed.